Skip to Main content Skip to Navigation
Conference papers

Privacy-Preserving $k$-means Clustering: an Application to Driving Style Recognition

Abstract : With the advent of connected vehicles, drivers will communicate personal information describing their driving style to their vehicles manufacturers, stakeholders or insurers. These information will serve to evaluate remotely vehicle state via an e-diagnostics service, to provide over-the-air update of vehicles controllers and to offer new third parties services targeting profiled drivers. An inherent problem to all the previous services is privacy. Indeed, the providers of these services will need access to sensitive data in order to propose in return an adequate service. In this paper, we propose a privacy-preserving $k$-means clustering for drivers subscribed to the pay how you drive service, where vehicles insurance fees are adjusted according to driving behavior. Our proposal relies on secure multi-party computation and additive homomorphic encryption schemes to ensure the confidentiality of drivers data during clustering and classification.
Complete list of metadata
Contributor : Witold Klaudel Connect in order to contact the contributor
Submitted on : Monday, August 24, 2020 - 7:11:32 PM
Last modification on : Sunday, June 26, 2022 - 12:30:52 AM




Othmane El Omri, Aymen Boudguiga, Malika Izabachène, Witold Klaudel. Privacy-Preserving $k$-means Clustering: an Application to Driving Style Recognition. NSS 2019 - 13th International Conference on Network and System Security, Dec 2019, Sapporo, Japan. pp.685-696, ⟨10.1007/978-3-030-36938-5_43⟩. ⟨hal-02920663⟩



Record views